Time-delayed linear quadratic optimal control problems
The book treats open- and closed-loop solutions of deterministic linear quadratic optimal control problems in which the state equation as well as the cost functional have time delays in the state and in the control. To include discrete as well as distributed delays, the delays are described by integrals with respect to Radon measures.\N\NThe control problem is set up with delays in \(\mathbb{R}^n\) and then ``lifted to a problem without delays in a suitable infinite-dimensional Hilbert space, leading to e new type of control operator which makes the problem different from standard infinite dimensional control problems. So the methods of solution of the latter which are found in the literature do not apply here, and a new approach is presented. The open-loop solvability of the lifted problem is now characterized by the solvability of a system of forward-backward integral evolution equations and the convexity of the cost functional. Returning to the underlying finite-dimensional problem, the adjoint equations involve coupled partial differential equations, again different from what is found in the literature. The closed-loop solvability is characterized by the solvability of three equivalent integral operator-valued Riccati equations, leading for the underlying finite-dimensional problem to coupled matrix-valued Riccati equations as well as coupled partial differential equations. Finally, the closed-loop representation of open-loop optimal control is achieved by three equivalent integral operator-valued Riccati equations. The proofs are detailed, and the presentation is largely self-contained.
This page was built for publication: Time-delayed linear quadratic optimal control problems
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7021899)
